TECTONICS, VOL. 22, NO. 4, 1028, doi:10.1029/2002TC001435, 2003
Mode of lithospheric extension: Conceptual models
from analogue modeling
Laurent Michon1
Netherlands Organization for Applied Scientific Research – Netherlands Institute of Applied Geoscience (TNO-NITG),
Utrecht, Netherlands
Olivier Merle2
Laboratoire Magmas et Volcans, Université de Clermont, Clermont-Ferrand, France
Received 5 July 2002; revised 8 November 2002; accepted 7 March 2003; published 9 July 2003.
[1] Comparison of analogue experiments at crustal
and lithospheric scale provides essential information
concerning the mode of deformation during
lithospheric extension. This study shows that during
extension, lithospheric deformation is controlled by the
development of shear zones in the ductile parts. At
lithospheric scale, the global deformation is initiated by
the rupture of the brittle mantle lithosphere. This failure
generates the formation of conjugate and opposite
shear zones in the lower crust and the ductile mantle
lithosphere. The analysis of the internal strain of the
ductile layers suggests that the two opposite shear
zones located below the asymmetric graben in the
lower crust and the ductile mantle lithosphere prevail.
Experiments show that from a similar initial stage, the
relative predominance of these shear zones originates
two different modes of deformation. If the crustal shear
zone prevails, a major detachment-like structure
crosscuts the whole lithosphere and controls its
thinning. In this model named the simple shear mode,
the resulting geometry shows that crustal and
lithospheric thinning are laterally shifted. If the
mantle shear zone predominates, the lithospheric
thinning is induced by the coeval activity of the two
main shear zones. This process called the necking
mode leads to the vertical superposition of crustal and
mantle lithospheric thinning. Applied to natural
laboratories (West European rift, Red Sea rift and
North Atlantic), this conceptual model allows a
plausible explanation of the different geometries and
evolutions described in these provinces. The North
Atlantic and the Red Sea rift systems may result from a
simple shear mode, whereas the necking mode may
explain part of the evolution of the West European
rift especially in the Massif Central and the Eger
1
Also at Geologisches Institut, Universität Freiburg, Freiburg, Germany.
Now at Laboratoire Magmas et Volcans, OPGC, Université Blaise
Pascal, CNRS, Clermont-Ferrand, France.
2
Copyright 2003 by the American Geophysical Union.
0278-7407/03/2002TC001435$12.00
graben.
INDEX TERMS: 8109 Tectonophysics: Continental
tectonics—extensional (0905); 1744 History of Geophysics:
Tectonophysics; 1645 Global Change: Solid Earth; 9335
Information Related to Geographic Region: Europe; KEYWORDS:
lithospheric extension, analogue modeling, West European rift, Red
Sea rift, passive margins. Citation: Michon, L., and O. Merle,
Mode of lithospheric extension: Conceptual models from analogue
modeling, Tectonics, 22(4), 1028, doi:10.1029/2002TC001435,
2003.
1. Introduction
[2] The asymmetric feature of continental rifts has been
clearly demonstrated during the past decades, whatever the
origin of rifting (i.e., passive or active rifting) [e.g., Mart
and Hall, 1984; Bosworth, 1985; Brun et al., 1992;
Rosendahl et al., 1992]. Although a single passive margin
presents an asymmetric geometry, conjugate and opposite
margins like in the northern Atlantic (the Canadian and
Iberian margins) mostly reveal a symmetric shape at the
lithospheric scale [Lister et al., 1991; Brun and Beslier,
1996]. Contrary to McKenzie [1978] and Wernicke [1981,
1985], who have respectively advocated pure shear and
simple shear to interpret lithospheric deformations,
Cochran and Martinez [1988], Lister et al. [1991] and
Kusznir and Ziegler [1992] have assumed a combination
of both simple and pure shear to explain these different
geometries. According to these authors, crustal deformation may be controlled by low angle detachment faulting,
and thinning of the mantle lithosphere may result from
pure shear. Such models have been improved by Beslier
[1991] and Brun and Beslier [1996] to explain the
occurrence of shear zones in the mantle lithosphere [Girardeau et al., 1988; Vissers et al., 1995]. These authors
have proposed that mantle lithospheric thinning could
result from a necking of the lithosphere, generated by
opposite and conjugate shear zones (Figure 5 of Brun and
Beslier [1996]).
[3] In the present paper, we first summarize the main
characteristics of several continental rifts and passive margins in order to show the different geometries resulting
from extension. We have selected a number of structures
which formed in different geodynamical contexts during the
Mesozoic or the Cenozoic: the West European rift, the
2-1
2-2
MICHON AND MERLE: MODE OF LITHOSPHERIC EXTENSION
Figure 1. Synthetic geological map of the West European rift showing the Massif Central rift, the Rhine
graben and the Eger graben. (a) Crustal cross section of the southern part of the Rhine graben showing the
asymmetry of the graben (modified after Brun et al. [1992]). The master fault bounds the graben in the
west and is probably connected to the Moho by a shear zone in the lower crust. (b) Crustal geometry of
the Massif Central rift. The mirror symmetry revealed by the graben geometry and the Moho inflexions
below the Forez and the Monts du Lyonnais results from symmetric extension between Priabonian and
middle Oligocene. The development of the syn-rift volcanism in the western part of the rift and the Moho
inflexion below the Limagne master fault are interpreted as the consequence of an asymmetric extension
during upper Oligocene and lower Miocene.
North Atlantic passive margins and the Red Sea rift.
According to the age of the continental lithosphere in each
province, we assume that the vertical strength profile
corresponds to a 4-layer model with two brittle high
strength levels (the upper crust and the upper mantle
lithosphere) and two ductile layers (the lower crust and
the lower mantle lithosphere). In such a context, it has been
suggested that the lithospheric deformation is initiated and
controlled by failure of the brittle mantle lithosphere [Brun
and Beslier, 1996].
[4] As the aim of this paper is to determine the process
leading to continental rifts at lithospheric scale, we compare scaled analogue experiments at crustal scale [Michon
and Merle, 2000] and lithospheric scale [Brun and Beslier,
1996]. In the 4-layer model, which represents the lithospheric strength profile, strain analysis of the ductile
layers makes possible to define two main shear zones,
which control the deformation and the thinning of the
models after the boudinage of the brittle mantle lithosphere. This study shows that the simple shear mode and
the necking of the lithosphere result from a similar initial
evolution. Finally, we propose a model that can explain
the various evolutions and geometries observed in the
West European rift, the North Atlantic passive margin and
the Red Sea rift.
2. Natural Laboratories
2.1. West European Rift
[5] The West European rift is composed of three provinces extending from the Bohemian Massif in the east,
through the Rhenish Province in the central part, to the
Massif Central in the west (Figure 1). In all provinces,
extension started contemporaneously 40 m.y. ago and led to
the formation of a set of grabens and a syn-rift volcanic
activity [Michon, 2001]. The global evolution with extension and sedimentation near sea level, closely followed in
the Massif Central and the Eger graben (Bohemian Massif)
by a volcanic activity, is interpreted in terms of passive
rifting evolution [Bois, 1993; Merle et al., 1998].
[6] In the Rhenish Province, the Rhine graben is a linear
structure which results from W/WNW-E/ESE extension.
Geophysical data clearly show that the Rhine graben is a
single asymmetric graben of 35– 40 km wide, which is
associated with strong crustal thinning [Brun et al., 1992].
MICHON AND MERLE: MODE OF LITHOSPHERIC EXTENSION
In the north, the half graben is bounded in the east by a west
dipping major fault. South to the Lalaye-Lubine-BadenBaden Variscan fault, the asymmetry is reversed and the
graben is limited in the west by an east dipping master fault
(Figure 1a). The distribution of the depocenters close to
these master faults reveals that the faults have controlled the
subsidence [Doebl and Olbrecht, 1974; Villemin et al.,
1986]. From the Priabonian to the end of the middle
Oligocene, a similar subsidence rate [Villemin et al., 1986]
and sedimentation at sea level [Sissingh, 1998] characterize
the northern and southern parts suggesting that the amount
of extension was identical at the graben scale. From the Late
Oligocene, subsidence drastically decreased in the southern
part whereas it was ongoing in the northern segment up to
Middle Miocene. After an interruption of sedimentation
during the Late Miocene, the northern Rhine graben was
reactivated at the beginning of the Pliocene [Sissingh,
1998]. Subsidence was still active in the Pleistocene in
the northern part and restarted in the southern part [Schumacher, 2002]. Whereas the Eocene-Oligocene rifting is
clearly related to the extension of the West European rift,
this latter extensional episode is closely associated with the
recent evolution of the Roer Valley graben and the southern
North Sea [Kooi et al., 1991; Zijerveld et al., 1992]. The
twofold evolution of the northern part of the Rhine graben
has resulted in a thicker pile of sediments and stronger
crustal thinning than in the southern part. In the northern
part of the Rhine graben, the stretching value deduced from
the geometry of the faults in the upper crust is estimated to
be in the range from 5 to 7 km. Assuming preservation of
the volume of the crust, a total extension value of about 17
km can be deduced [Brun et al., 1992]. In the southern part,
the extension value estimated from preservation of the
volume of the crust is about 12 km [Bois, 1993]. As it
has been shown that the northern segment was affected by
two distinct periods of extension, the extension value
resulting from the West European rift episode only is
probably about 12 km. Thus the extension rate deduced
from the 10 m.y. Priabonian-Middle Oligocene rifting
episode is close to 1.2 mm/yr.
[7] The Massif Central rift, which is the most important
segment of the West European rift, has been recently
reinterpreted [Merle et al., 1998; Michon, 2001]. The
Variscan basement is affected by a north-south 180 km
wide rift composed by Eo-Oligocene grabens. Along an
east-west cross section, the rift is characterized by two
lateral and opposite half grabens (the Limagne graben in the
west and the Bresse graben in the east) and a central and
near-symmetric graben (the Roanne-Montbrison graben)
(Figure 1b). Geophysical data show that the structure is
bounded by the two opposite detachment faults of the
Limagne and Bresse half grabens which might be
connected at depth to the Moho discontinuity [Michon,
2001]. This overall geometry reveals a strikingly mirror
symmetry on both sides of the central part of the RoanneMontbrison graben (Figure 1b). The east-west cross section
also shows asymmetrical features at the scale of the rift.
The Moho inflexion below the Limagne half graben and the
occurrence of a magmatic activity restricted in the western
2-3
part of the rift correspond to geological characteristics
which are hardly compatible with a purely symmetric
evolution. The explanation can be found in the sedimentary
record [Merle et al., 1998]. Similar deposits in the Limagne
and Bresse half grabens from the Late Eocene to the Early
Oligocene together with sedimentation at sea level [Giraud,
1902; Rat, 1974; Bodergat et al., 1999] suggest a symmetric extension during this period. However, in the Late
Oligocene, persistence of marine incursions in the Limagne
half graben and the lack of marine incursion and the lull in
sediment deposits in the Bresse graben show that the
overall symmetrical evolution came to a halt. During this
period, crustal thinning occurred in the western Limagne
graben. This is clearly established from geophysical profiles
where crustal thinning reaches 25% in the Limagne graben
whereas it does not exceed 13% in the Bresse graben. The
asymmetric extension during the Late Oligocene is confirmed by the occurrence of a scattered magmatic activity in
the Limagne half graben. Then, we interpret the mirror
symmetry as resulting from the Late Eocene-Middle Oligocene symmetric extension whereas the Late OligoceneEarly Miocene evolution corresponds to an asymmetric
extension located in the western part of the rift only. The
stretching value deduced from master fault geometries in
the upper crust is inferior to 10– 15 km. However, as in the
Rhine graben, this value is more than two times lower than
the value estimated from geophysical data, which is close to
25– 30 km [Bois, 1993]. Again, this discrepancy may be
explained by ductile flow in the lower crust. The extension
rate estimated for the Massif Central rift ranges from 2.5 to
3 mm/yr.
[8] In the Bohemian Massif, the evolution and the
geometry of the Eger graben are still poorly constrained.
This graben is 40– 45 km wide, trending N60E, and is
bounded in the north by a main normal fault (the Krušné
Hory fault), which gives to the whole structure a global
asymmetry. The Priabonian age of the oldest sediments
related to the extension [Chlupac et al., 1984] indicates
that the graben formation was coeval with the formation of
the Massif Central rift and the Rhine graben. The thickness
of the Late Oligocene-Middle Miocene sediments never
exceeds 500 –600 m and the depocenter are located close
to the Krušné Hory fault suggesting (1) a small amount of
extension and (2) a major role of this master fault in the
graben evolution. From the Middle Oligocene (i.e., 10 m.y.
ago after the onset of the extension) and up to the Middle
Miocene, a syn-rift volcanic activity was located within the
graben leading to the formation of large magmatic provinces
[Ulrych et al., 2000]. Despite the few amount of data, the
formation of a single asymmetric graben and the development of a syn-rift volcanic phase within the extensional area
correspond to geological information of primary importance, which can be help to determine the evolution and
the deformation of the whole lithosphere.
2.2. North Atlantic Passive Margins
[9] The understanding of the North Atlantic passive
margins has been largely improved by the discovery of a
2-4
MICHON AND MERLE: MODE OF LITHOSPHERIC EXTENSION
Figure 2. Reconstruction of the North Atlantic passive margin before oceanization (after Tankard and
Welsink [1987]). In this cross section, only structures resulting from the main period of extension are
represented (see text for explanation). 1: Galicia bank. 2: Interior Basin. JAB: Jeanne d’Arc Basin. FC:
Flemish Cap. DGM: Deep Galicia Margin.
mantle ridge close to the ocean-continent transition zone
in the Galicia passive margin [Boillot et al., 1980]. At
the latitude of Iberia, geophysical and tectonic studies
have allowed to reconstruct the initial geometry of the rift
before oceanization taking into account both Galicia and
East Canadian margins (Figure 2) [Tankard and Welsink,
1987]. Seismic data show that the East Canadian margin is
composed from west to east of a nearly symmetric graben
(the Jeanne d’Arc basin) and a horst (the Flemish cap). The
subsidence analysis of the Jeanne d’Arc basin reveals a
main period of quick subsidence during the Late Barremian
(around 120 Ma) [Driscoll et al., 1995]. In the east, the
Iberian margin can be divided in two different parts: the
Interior Basins (Interior and Porto basins) and the Deep
Galicia margin. The Interior Basins correspond to symmetric grabens of the proximal part of the margin which were
active from the very beginning of the extension (i.e.,
Valanginian) [Manatschal and Bernoulli, 1999]. In contrast,
the Deep Galicia margin is composed of eastward tilted
blocks which yield a strongly asymmetric structure. The age
of the syn-rift sediments (Hauterivian to Aptian) suggests
that the extension has shifted from the Interior Basins to the
Deep Galicia margin during rifting [Manatschal and Bernoulli, 1999]. Stretching values have been deduced for the
whole period of rifting (300– 350 km) [Keen and Dehler,
1993]. However, the multiphase extension in the North
Atlantic leading to continental breakup as well as the
previous extension in the Interior Basin [Manatschal and
Bernoulli, 1999] make it difficult the estimation of the
stretching value related to the main phase of rifting (Hauterivian to Albian).
[10] One of the main characteristics of the Galicia
passive margin is the presence of a prominent seismic
reflector, the so-called S reflector. In the Deep Galicia
margin, the S reflector is located at the base of the tilted
blocks and its dip evoluates from westward dipping in the
east to a slightly eastward dipping close to the mantle
ridge [Boillot et al., 1980, 1988; Mauffret and Montadert,
1987]. In this area, the S reflector directly overlies the
mantle ridge and seems to have merged into seafloor.
Whereas the S reflector is mainly interpreted as a top-tothe-ocean low-angle detachment structure which was
responsible for mantle exhumation during rifting [e.g.,
Beslier et al., 1990; Reston et al., 1996; Manatschal and
Bernoulli, 1999], the origin and the age of a second
opposite shear zone (top-to-the-continent) in the mantle
peridotite is still a matter of controversy. Boillot et al.
[1995] and Brun and Beslier [1996] have proposed to
explain the mantle exhumation by the development of two
nearly synchronous and opposite shear zones in both
the lower crust and the mantle lithosphere leading to the
necking of the lithosphere. According to this model, the S
reflector and the mantle shear zone could correspond to
the detachment structures in the lower crust and the
mantle, respectively. However, using 40Ar/39Ar ages [Féraud et al., 1988], Manatschal and Bernoulli [1999] have
shown that the tectonic activity of the mantle shear zone is
slightly older than the crustal one and have proposed that
both mantle and crustal structures result from two different
stages in continental extension. Whatever the exact chronology of these two detachment faults, the evolution of
the syn-rift subsidence in both East Canadian and Galicia
MICHON AND MERLE: MODE OF LITHOSPHERIC EXTENSION
Figure 3. Geological map of the Red Sea area showing the
distribution of the syn-rift volcanism on the Arabian
microplate only.
opposite margins suggests that (1) maximal crustal thinning was superimposed on the two graben areas (Jeanne
d’Arc basin and Deep Galicia margin) and that (2) mantle
lithospheric thinning took place in the eastern part of the
rift (i.e., east to the Flemish Cap) (Figure 12 of Keen and
Dehler [1993]).
2.3. Red Sea Rift
[11] The Red Sea rift is a 1700 km length NNW-SSE
oriented narrow rift which is located north to the Afar area
(Figure 3). During the last decades, this rift was the
subject of two main controversies. On one hand, the
debate focuses on the origin of the extension (i.e., passive
rifting versus active rifting) and the relation with the Afar
plume [Courtillot, 1980; Dixon et al., 1989; Bohannon et
al., 1989]. Recently, Menzies et al. [1997] have proposed
that the extension in the northern Red Sea results from the
northward propagation of the southern Red Sea, which is
linked to the Afar plume. In their model, an active rifting
origin is clearly advocated for the southern Red Sea,
whereas the evolution of the northern segment can be
interpreted in terms of passive rifting. On the other hand,
assuming that the Red Sea rift results from a passive
2-5
rifting evolution, two opposite modes of extension have
been proposed to explain the main geological and geophysical features of the area. Wernicke [1985] has referred
to the development of a low-angle detachment structure
(i.e., simple shear mode) to induce volcanism and uplift in
the eastern border of the rift only (in the Saudi Arabian
microplate). In contrast, Buck et al. [1988] have advocated
a pure shear mode with an initial 110 km wide rifting area
becoming narrower with time (20 km wide) to explain the
present-day high and narrow heat flow in the central Red
Sea. It is well worth noting that the simple shear mode is
based on Eo-Miocene geological features, whereas presentday data are used to define the pure shear model. Consequently, these two modes of extension are deduced from
different periods: the simple shear mode for an initial
continental extension and the pure shear mode for a nearly
oceanization stage.
[12] As the southern part of the Red Sea rift presents
strong interactions with the Afar plume, we will summarize the main Oligo-Miocene geological features of the
northern Red Sea in order to constrain the evolution of the
rifting. Although the onset of the extension is not well
established, Bohanonn et al. [1989] have suggested a
normal faulting activity with sedimentation close to sea
level from 23 to 29 Ma, with an increased extension rate
around 25 Ma. During this period a first phase of
volcanism started in the Arabian margin only. The age
of the onset of magmatism is poorly constrained. K/Ar
ages mainly obtained in the 1970s and 1980s suggest the
occurrence of volcanism around 30– 32 Ma (i.e., before
extension) [Brown et al., 1984]. However, several combined K/Ar and 40Ar/39Ar data have often shown strikingly
differences between ages obtained with both methods
suggesting that ages based on the K/Ar method must be
handled with care [Baker et al., 1996; Rittmann and
Lippolt, 1998]. The oldest lava flows dated from the
40
Ar/39Ar method in Saudi Arabia indicate an onset of
the volcanic activity coeval with the beginning of the
extension (27–28 Ma) [Chazot et al., 1998]. After a lull
of volcanism during several millions years, the Arabian
margin was again affected by an intense magmatic phase
coeval with a main period of continental rifting between
21– 24 Ma [Chazot et al., 1998]. This magmatism led to
the formation of a 1000 km length dyke swarm parallel to
the newly formed continental rift. Finally, from the Early
Miocene (ca. 20 Ma), the Arabian margin was uplifted
causing the present-day topographic asymmetry on either
side of the Red Sea [Bohannon et al., 1989]. According to
many authors [e.g., Wernicke, 1985; Bohannon et al.,
1989], we consider that this general evolution of the
period of continental rifting (i.e., rifting at sea level,
volcanism and uplift) can be interpreted in terms of
passive rifting. The overall geometry resulting from this
Oligo-Miocene event reveals a strong asymmetry of the
spatial distribution of magmatism and uplift restricted to
the Arabian microplate. Likewise, this asymmetry can be
readily seen in the seafloor topography of the Northern
Red Sea (Figure 5 of Cochran and Martinez [1988]). The
bathymetry decreases abruptly between the western shore
2-6
MICHON AND MERLE: MODE OF LITHOSPHERIC EXTENSION
experiments resulting from independent studies, we first
develop a common scaling approach defining several dimensionless numbers for the crustal and lithospheric experiments, in order to show that these experiments are properly
scaled.
3.1. Experimental Background
Figure 4. Strength profile (a) for a continental lithosphere
with a normal thermal gradient and (b) in experiments. Sand
is used to simulate the brittle parts of the lithosphere and
silicone putty mimics the behavior of the lower crust and the
ductile mantle lithosphere.
and the central axis of the Red Sea, whereas it gradually
increases from the axial zone up to the eastern border.
Contrary to the Oligo-Miocene asymmetric features, the
present-day heat flow and seismicity distribution reveals a
symmetric geometry [Cochran and Martinez, 1988; Dixon
et al., 1989], which could indicate a change in the rifting
evolution.
[13] Although the extension rate of the northern Red Sea
is fairly well estimated from the Late Miocene to the
present-day (1cm/yr) [Cochran and Martinez, 1988], the
stretching values and the extension rate are not known for
the pre-Late Miocene period. Assuming that most of the
extension has occurred in the past 12– 14 Ma [Le Pichon
and Gaulier, 1988], the extension rate may have been very
slow during the initial period of rifting.
3. Analogue Experiments
of Continental Extension
[14] In this part, we compare analogue experiments already
published and carried out at crustal [Michon and Merle, 2000]
and lithospheric scale [Brun and Beslier, 1996]. These
[15] The analogue models can be compared to the
natural systems if the distribution of stresses, densities
and rheologies are similar in nature and experiments
[Hubbert, 1937]. In the areas presented in the first
sections of this paper, the lithosphere is assumed to have
a normal thermal gradient before the rifting event. It is
generally accepted [e.g., Ranalli and Murphy, 1987; Buck,
1991; Davy and Cobbold, 1991; Fernandez and Ranalli,
1997; Burov and Poliakov, 2001] that with such a thermal
gradient (Moho-temperature less than 500– 600C), the
strength profile of the lithosphere before the rifting process is characterized from top to bottom by a brittle upper
crust, a ductile lower crust, a brittle mantle lithosphere
and a ductile mantle lithosphere (Figure 4a). The rheological structure of the lithosphere being highly dependent
on the thermal gradient and the strain rate, this theoretical
4-layer strength profile can change during the thinning of
the lithosphere with a shift of the brittle-ductile transition in
the crust and the mantle lithosphere [Davy and Cobbold,
1991]. Experiments at lithospheric scale have been carried
out with a 4-layer brittle-ductile model whereas experiments at crustal scale correspond to the upper part of the
strength profile only (Figure 4b). In both experiments, the
ductile and brittle behaviors of the lithosphere were simulated by silicone putty and sand layers, respectively. Although the behavior of the lithospheric ductile levels is
characterized by a power law rheology, Davy and Cobbold
[1991] have shown that the use of the silicone putty is a
good approximation to simulate the lithospheric ductile
parts.
[16] Numerical approaches show that the thermal regime
of the lithosphere could play a major role on the mode of
deformation during a rifting event [e.g., Buck, 1991; Burov
and Poliakov, 2001]. The influence of the temperature
Table 1. Geometric and Mechanical Variables in Nature and Experiments at Crustal and Lithospheric Scale
Experiments
Variable
hUBL
hUDL
Hc
HML
rUBL
rUDL
rLBL
rLDL
m
t0
e
g
Definition
thickness of the upper crust (UBL)
thickness of the lower crust (UDL)
thickness of the crust (UBL + UDL)
thickness of the mantle lithosphere (LBL + LDL)
density of the upper crust (UBL)
density of the lower crust (UDL)
density of the brittle mantle lithosphere (LBL)
density of the ductile mantle lithosphere (LDL)
angle of internal friction
viscosity
cohesion
strain rate
gravity acceleration
Nature
Crustal Scale
3
15 – 18 10
12 – 15 103
30 – 35 103
70 – 90 103
2700
2800
3300
3300
30
1 1021
2 106
5 1015
10
2
1.5 10
1.5 102
3 102
1400
1350
30
3 104
1
1.8 104
10
Lithospheric Scale
2
1.2 – 1.4 10
0.5 – 0.6 102
1.7 – 2 102
2.3 – 2.8 102
1400
1100 – 1214
1400
1305
30
1.3 – 2.1 104
1
2.3 104 – 2.3 103
10
Dimensions
m
m
m
m
kg/m3
kg/m3
kg/m3
kg/m3
Pa.s
Pa
s1
m/s2
2-7
MICHON AND MERLE: MODE OF LITHOSPHERIC EXTENSION
Table 2. Dimensionless Numbers in Nature and Experiments
Experiments
Dimensionless
Parameter
Definition
Nature
Crustal Scale
Lithospheric Scale
1
1b
2
3
3b
4
5
6
7
UBL/UDL thickness
crust/mantle lithosphere thickness
angle of internal friction
UBL/UDL density
LBL/LDL density
gravitational/viscous stresses
inertial/viscous stresses
failure resistance/viscous stresses
gravitational stress/cohesion
1 – 1.5
0.4
30
0.96
1
81
3 1024
31.75
202.5
1
30
1.04
37.5
2.9 1011
15.16
210
2.3 – 3
0.6 – 0.85
30
1.12 – 1.27
1.07
4 – 34.5
9
5.6 10 – 1.57 1011
2.1 – 16.08
168 – 196
cannot be included directly in the analogue experiments.
Nevertheless, in the experiments, a change of the thickness
ratio between the brittle and ductile layers (i.e., thick upper
crust and thin lower crust or thin upper crust and thick lower
crust) is able to simulate different initial geotherms from an
old and cold lithosphere up to a young and hot lithosphere.
In addition, Michon and Merle [2000] have shown that the
initial thermal regime of the lithosphere is a parameter that
has a minor influence on the crustal deformation. Then, we
consider that for the study of the first stage of a rifting
process, analogue approach can be used to study the mode
of deformation of the lithosphere.
[17] To guarantee similarity in nature and experiments,
we define several variables related to the geometry and
the properties of the materials (Table 1). The geometry of
the different parts of the lithosphere is characterized by the
thickness of the Upper Brittle Layer hUBL (i.e., the upper
crust in nature), the Upper Ductile Layer hUDL (i.e., the
lower crust), the whole crust HC and the mantle lithosphere
HML. Materials property variables are the density of the
Upper Brittle Layer rUBL, of the Upper Ductile Layer rUDL,
of the Lower Brittle Layer rLBL and of the Lower Ductile
Layer rLDL. The viscosity m of the ductile layer, and the
internal angle of friction and the cohesion to of the brittle
levels are also variables related to the materials properties.
The relation between the deformation and the time corresponds to the strain rate e. Finally, the only force is the
gravity acceleration g.
[18] According to the Buckingham- theorem, there are
10 variables minus 3 dimensions equal to 7 independent
dimensionless numbers (Table 2) that need to maintain the
same value in nature and experiments. Of these, two
dimensionless parameters correspond to the geometric ratios
of the system:
Y
1
Y
1b
crust in nature) and the Upper Ductile Layer UDL of the
system (i.e., the ductile lower crust), respectively. The
second dimensionless parameter 1b is only used to scale
the lithospheric experiments and corresponds to the thickness ratio between the crustal HC and mantle lithospheric
HML levels.
[20] The third dimensionless number can be defined as
the frictional angle of the brittle material:
Y
2
HC
¼
:
HML
Y
3
Y
3b
Y
rUBL
;
rUDL
ð4Þ
¼
rLBL
:
rLDL
ð5Þ
¼
rghUBL
;
em
ð6Þ
¼
erh2UBL
;
m
ð7Þ
2 tan rghUBL
þ tan :
3em
ð8Þ
4
5
Y
[19] In equation (1), hUBL and hUDL represent the thickness of the Upper Brittle Layer UBL (i.e., the brittle upper
¼
[21] 3 defines the density ratio between the UBL and
the UDL, whereas 3b allows to scale the density of the
Lower Brittle Layers LBL and Lower Ductile Layers LDL
in the experiments at lithospheric scale only.
[22] Knowing that gravity is balanced in the brittle
materials by stresses that resist failure and in the ductile
levels by inertial and viscous stresses, we define three other
dimensionless parameters, which correspond to ratios of
these stresses:
ð1Þ
ð2Þ
ð3Þ
Two other dimensionless parameters must be related to the
density ratio of the different layers.
Y
hUBL
¼
;
hUDL
¼ :
6
¼
[23] Equations (6), (7) and (8) represent respectively
ratios between the gravitational stress and the viscous stress,
the inertial stress and the viscous stress (i.e., the Reynolds
2-8
MICHON AND MERLE: MODE OF LITHOSPHERIC EXTENSION
Figure 5. Experimental setup used in analogue experiments at (a) crustal and (b) lithospheric scale.
Apparatus at lithospheric scale after Brun and Beslier [1996].
number) and the failure resistance stress and the viscous
stress (see Merle and Borgia [1996] for details).
[24] Finally, the last dimensionless parameter corresponds to the ratio of gravitational stress to cohesion:
Y
7
¼
rghUBL
t0
ð9Þ
Results of equations (1), (1b), (3) and (3b) clearly show
the similarity between nature and the experiments at both
scales for the geometry and the density of the materials
(Table 2). The angle of internal friction is the same for the
sand and natural rocks (30). Considering an average
strain rate in nature (5 1015 s1) and a viscosity of
the lower crust of m = 1021 Pa s, 4 and 6 impose a
strain rate in the experiments around 104 –103 s1. The
very small value of 5 in nature (3.1 1024) shows that
the inertial forces are negligible with respect to the viscous
forces. In the experiments, such a low value cannot be
reached even using materials with high viscosity and low
density. Nevertheless, in experiments the low values of 5
(5.6 109 – 1.57 1011) suggest that the inertial forces
have also a minor importance with respect to the viscous
forces. Thus this number may receive no further consideration. The last dimensionless parameter 7 is partly
derived from the cohesion of the materials. The cohesion
of intact and massive rock is considered to be 107 Pa
[Jaeger and Cook, 1971]. However, it is widely accepted
that the cohesion could decrease of 1 or 2 orders of
magnitude for fractured rocks [e.g., Merle et al., 2001].
In nature, the continental crust is generally affected by
numerous faults resulting from different tectonic events
(e.g. the Variscan continental crust in Europe). Then, it is
reasonable to consider that the cohesion of the crust at a
large scale is closer to 106 Pa than 107 Pa (Table 1). The
resulting 7 is similar in nature and in experiments at
crustal and lithospheric scale (Table 2).
[25] This scaling approach finally shows that (1) experiments carried out by independent studies [Beslier, 1991;
Brun, 1999; Michon and Merle, 2000] can be compared and
(2) the results and interpretation can be applied to natural
laboratories.
[26] It has been proposed that the deformation in a
lithosphere is initiated by failure in the brittle mantle
lithosphere [Allemand et al., 1989; Brun and Beslier,
1996]. In numerical modeling, such a failure is usually
simulated by a preexisting weak zone in the lithosphere
[e.g., Huismans and Beaumont, 2002]. In the analogue
experiments, the rupture in the brittle mantle lithosphere
was initiated by a velocity discontinuity (VD), which was
forced by the border of a moving plastic sheet (Figure 5). As
in the numerical approach, the VD concentrates the deformation and controls the development of the structures. At
crustal scale, the VD was located at the base of the system
(i.e., below the silicone layer corresponding to the lower
crust), whereas it was situated on the vertical lateral boundaries of the apparatus in the experiments at lithospheric
scale (Figure 5). Contrary to the analogue modeling at
lithospheric scale, Michon and Merle [2000] have carried
out experiments with one but also two VDs in order to
determine the role of the number of ruptures upon crustal
deformation. In experiments with one VD, the use of a
vertically stratified silicone has allowed the analysis of the
internal strain within the ductile layers and the determination of the highest strain zones (i.e., the shear zones).
3.2. Summary of Previous Studies
[27] Analogue experiments of continental rifting have
been conducted during the last decade at crustal [Michon
and Merle, 2000] and lithospheric scale [Beslier, 1991;
Brun and Beslier, 1996; Chemenda et al., 2002]. These
studies have shown that major parameters control the mode
of deformation and the strain during a rifting process: the
extension rate and the development of shear zone in the
mantle lithosphere and the crust. We summarize herein
the main observations.
[28] Experiments with one VD and various extension
rates show in both studies the occurrence of two deformation styles distinguished by the number of grabens. Low
extension rate leads to the formation of a single half-graben
on the movable part (Figure 6a). At both crustal and
lithospheric scales, this asymmetric graben is bounded by
a master fault in the brittle upper part, which is rooted along
a slight silicone upwelling. In contrast, high extension rate
MICHON AND MERLE: MODE OF LITHOSPHERIC EXTENSION
2-9
Figure 6. Comparison between experiments carried out at crustal and lithospheric scale. (a) At low
extension rate the extension induces the formation of a single asymmetric graben on the movable part. (b)
In contrast, a high extension rate leads to the creation of two grabens: a near-symmetric graben in the
motionless part and an asymmetric graben in the movable part. Experiments at lithospheric scale after
Beslier [1991].
generates simultaneously the formation of two parallel and
linear grabens separated by a horst: a half graben on the
movable part and a near-symmetric graben on the motionless part (Figure 6b). At lithospheric scale, it is worth noting
that the maximum value of thinning of the analogue crustal
layers is located in the central part of the asymmetric
graben, whereas the analogue layers of the mantle lithosphere are mostly thinned below the border of the horst. The
difference observed between the geometry of the nearsymmetric and asymmetric grabens in crustal and lithospheric experiments is likely to result from the uplift of the
UBL due the activity of the detachment fault and the
resulting isostatic adjustment in lithospheric experiments.
Whatever the extension rate, the deformation of the UBL
seems to be directly linked at lithospheric scale to the
rupture of the LBL or at crustal scale to the velocity
discontinuity. These results indicate that a low extension
rate prevents the formation of the symmetric companion
graben observed in high extension rate experiments.
[29] Using a vertically stratified silicone, the internal
strain of the ductile layers has also been studied at lithospheric scale in high extension rate experiments [Brun and
Beslier, 1996]. According to these authors, extension induces the formation of two conjugate shear zones in each
ductile part of the models (i.e., the LDL and the UDL),
which control the thinning of the whole lithospheric model
(see Figure 5 of Brun and Beslier [1996]). They have
proposed that the mode of lithospheric extension corresponds to a necking of the lithosphere that can lead to
mantle exhumation just before oceanization [Brun and
Beslier, 1996]. It is interesting to note that Chemenda et
al. [2002] have also determined the occurrence of litho-
spheric shear zones controlling the thinning of the models.
Nevertheless, their experiments show that for an evoluate
stage of extension only one lithospheric shear zone prevails
mimicking the simple shear mode proposed by Wernicke
[1981, 1985] (see Figure 5 of Chemenda et al. [2002]).
Such a result is in disagreement with the necking model
Figure 7. Internal strain of the silicone layers in high
extension rate experiments deduced from the final geometry
of the vertical markers in the silicone layers. Colors from
light to dark blue indicate variations in strain intensity for
the top-to-the-left shear zone whereas the yellow to red
colors indicate strain intensity for the top-to-the-right shear
zone (see text for explanation). See color version of this
figure at back of this issue.
2 - 10
MICHON AND MERLE: MODE OF LITHOSPHERIC EXTENSION
Figure 8. Cross section of an experiment at lithospheric
scale characterized by a high extension rate and a nearly
similar amount of extension (see text for explanation)
(modified after Brun and Beslier [1996]).
proposed by Brun and Beslier [1996] where shear zone
activity is coeval on both sides of the rift up to the formation
of the ocean lithosphere.
3.3. New Results From the Comparison
of Analogue Experiments at Crustal
and Lithospheric Scale
[30] To solve the apparently different deformation at
lithospheric scale, we compare the internal strain in the
experiments at crustal and lithospheric scales.
[31] High extension rate experiments at crustal scale
reveal that a VD always creates two conjugate shear zones
in the ductile level, which are clearly related to the grabens
in the UBL (see Figure 4 of Michon and Merle [2000]).
Shear measurements also indicate that the top-to-the-left
shear zone associated with the half graben always prevails
with respect to the other one (Figure 7a). In contrast, the
opposite shear zone, which is related to the near-symmetric
graben is much less important and its activity decreases
during extension.
[32] Using the same analytical approach in experiments at
high extension rate, we have determined the internal strain
of the silicone layers at lithospheric scale (Figure 7b). This
clearly shows the occurrence of a continuous shear zone
associated with the asymmetric graben which crosscuts the
whole model down to the LDL. Such a shear zone is
strikingly similar to the one observed in the experiments
of Chemenda et al. [2002]. Very high shear strain is
recorded in the UDL whereas strain is lower in the LDL.
This major shear zone corresponds to a top-to-the-left lowangle detachment structure which controls the upwelling of
the LBL. To this respect, it is interesting to note that the
asymmetric shape of the glucose solution (i.e., the analogue
asthenosphere) suggests that the detachment fault controls
the thinning of the whole model. It is also possible to define
two top-to-the-right shear zones that can be traced out in the
UDL and the LDL (Figure 7a). Shear measurements
deduced from the distortion of the initially vertical markers
indicate that the highest shear strain along the shear zones
are located below the asymmetric graben, that is in the UDL
for the top-to the-left shear zone and in the LDL for the topto-the-right shear zone. In a way similar to crustal scale
experiments, the deformation concentrates along the major
detachment fault and the activity of the top-to-the-right
shear zone in the UDL progressively decreases as it is
suggested by the preservation of the initial geometry of the
nearly symmetric graben.
[33] A second experiment at high extension rate (Figure 4a
of Brun and Beslier [1996]) shows a different geometry at
lithospheric scale. This model, conducted at an identical
extension rate that the previous experiment (i.e., 5 cm/h), has
been stopped at a slightly higher amount of extension (3.8 cm
instead of 3.2 cm for the first experiment). The lack of
vertically stratified silicone in the model does not allow the
internal strain to be defined. Nevertheless, the geometry of
the LBL with its uplift toward the nearly symmetric graben
suggests that the deformation is controlled by the top-to-theright shear zone located in the LDL (Figure 8). Crustal
deformation is characterized by the development of a
strongly asymmetric graben due to a top-to-the-left main
shear zone in the UDL similar to the main shear zone
observed in previous experiments. This top-to-the-left shear
zone was passively deformed when the top-to-the-right shear
zone became prominent (Figure 8). Thus, the total thinning
of the model (i.e., mantle levels and crustal layers) is
controlled by two conjugate and opposite main shear zones.
[34] In experiments at low extension rate, the internal
strain of the silicone layers has not been studied in details.
Nevertheless, the similar location and shape of the thinning
of the silicone layer in crustal scale experiments also
suggests the occurrence of two conjugate shear zones
(Figure 9). We speculate that the top-to-the right shear zone
in low rate experiments is not strong enough to induce the
formation of a nearly symmetric graben in the UBL.
[35] All experiments suggest that a VD creates two
opposite and conjugate shear zones in the ductile layers
whatever the extension rate. This result is in agreement with
purely brittle models in which a VD always leads to the
Figure 9. Comparison of the final geometry of the silicone
layer in experiment at crustal scale with (a) high and (b) low
extension rate. The similar shape of the thinning of the
silicone layer suggests the occurrence of two conjugate
shear zones in the experiments whatever the extension rate.
MICHON AND MERLE: MODE OF LITHOSPHERIC EXTENSION
2 - 11
Figure 10. Synthetic evolution of the lithospheric deformation during continental rifting at high
extension rates. During the initial stage of rifting, the extension induces (1) the development of conjugate
shear zones in the ductile layers and (2) the formation of two incipient grabens. The evoluate stage is
characterized by the predominance of two opposite shear zones located on the same vertical section. The
crustal shear zone lead to the development of the asymmetric graben. When the brittle mantle lithosphere
is broken out, the system can display two different mode of extension. In the necking model, the crustal
shear zones control the crustal thinning and the mantle shear zone located below the asymmetric graben
leads to the deformation of the mantle lithosphere. In the simple shear mode, the thinning and the
deformation of the whole lithosphere is controlled by the shear zone associated with the asymmetric
graben. V represents the potential location of the volcanism.
formation of two conjugate normal faults [e.g., Faugère et
al., 1986], and with the models of Chemenda et al. [2002]
where the lithosphere is modeled by an elasto-plastic
material. It also shows that the role of the top-to-the-right
shear zone increases as a function of the extension rate.
[36] To sum up, analysis of the internal strain in the
silicone layers shows (1) the development of two conjugate
shear zones in each ductile level (i.e., the UDL and LDL)
and (2) that the two shear zones located below the asymmetric graben predominate. Deformation of the mantle
lithosphere levels in a final stage (i.e., after the boudinage
of the LBL) is controlled by the shear zone located either in
the UDL (Figure 7) or in the LDL (Figure 8).
4. Interpretation and Discussion
4.1. Modes of Lithospheric Extension During
Continental Rifting
[37] Although analogue models at crustal and lithospheric
scale have been carried out using different apparatus, the
experimental results show striking similarities, proving that
modeled deformation is indicative of a common tectonic
process. As it has already been suggested [Beslier, 1991],
we interpret the final geometry of crustal layers as resulting
from the rupture of the brittle mantle.
[38] In high extension rate experiments, this rupture induces at the initial stage of extension a mirror symmetric
deformation characterized by conjugate shear zones on both
sides of the brittle mantle lithosphere (i.e., in the lower crust
and the mantle ductile lithosphere) (Figure 10a). During this
stage, the shear zones in the lower crust are responsible for the
creation of two incipient grabens in the upper crust. This
mode of extension mimics a pure shear deformation. With
increasing amount of extension, this symmetric evolution
stops. The activity of two shear zones is enhanced and they
become prominent with respect to the two others. They are
located on the same vertical section, one in the UDL and the
other in the LDL (Figure 10b). The shear zone, which prevails
in the UDL, induces the formation of an asymmetric graben
(Figure 10b). In nature, this evolution is thought to originate
the formation of an asymmetric graben (i.e. a half-graben)
associated with a master normal fault in the upper brittle crust.
[39] Experiments at lithospheric scale show two different
final geometries for a nearly similar amount of extension
[Brun and Beslier, 1996]. These experiments reveal that
during a later stage, the boudinage of the brittle mantle
lithosphere leads to two contrasting evolutions resulting
from the relative role of the main shear zones located in
the lower crust and the ductile mantle lithosphere. If the
UDL shear zone associated with the asymmetric graben
prevails with respect to the mantle shear zone, boudinage of
the brittle mantle lithosphere induces the formation of a
master shear zone (i.e., a low-angle detachment structure),
which crosscuts and controls the thinning of the whole
lithosphere (Figure 10c left). In contrast, when the top-to-
2 - 12
MICHON AND MERLE: MODE OF LITHOSPHERIC EXTENSION
Figure 11. Synthetic evolution of the lithospheric deformation during continental rifting at low
extension rates. During the initial stage of rifting, the extension induces (1) the development of conjugate
shear zones in the ductile layers and (2) the formation of one incipient grabens. The evoluate stage is
characterized by the predominance of two opposite shear zones located on the same vertical section. The
crustal shear zone lead to the development of the asymmetric graben. When the brittle mantle lithosphere
is broken out, the system can display two different mode of extension. In the necking model, the crustal
shear zones control the crustal thinning and the mantle shear zone located below the asymmetric graben
leads to the deformation of the mantle lithosphere. In the simple shear mode, the thinning and the
deformation of the whole lithosphere is controlled by the shear zone associated with the asymmetric
graben. V represents the potential location of the volcanism.
the-right shear zone located in the ductile mantle lithosphere
is connected with the UDL shear zone associated with the
nearly symmetric graben, the thinning of the lithosphere
mainly results from the coeval activity of the two opposite
and main shear zones (Figure 10c right). Thus, starting from
a similar geometry before the boudinage of the brittle
mantle lithosphere, extension can induce two different
modes of extension: a simple shear mode which presents
real similarities with the model proposed by Wernicke
[1985], or a necking of the lithosphere as it has been
described by Brun and Beslier [1996].
[40] In low extension rate experiments, it is assumed that
during the initial stage, rupture of the brittle mantle lithosphere induces the formation of two conjugate shear zones
in each ductile layer. However, experiments show that only
one UDL shear zone is strong enough to initiate a graben in
the upper crust. The resulting geometry defined by the set of
shear zones corresponds to a mirror symmetry on both
sides of the brittle mantle lithosphere (Figure 11a). As in
the high extension rate evolution, subsequent deformation
is controlled by the two main and opposite shear zones
located in the ductile mantle lithosphere and the lower crust
(Figure 11b). If the shear zone associated with the asymmetric graben prevails with respect to the mantle shear zone,
a single low-angle detachment structure crosscuts the whole
lithosphere when the brittle mantle lithosphere is broken
out. The mode of extension then corresponds to a simple
shear mode and the maximum thinning of the crustal layers
and the mantle lithospheric levels are not located on the
same vertical line (Figure 11c left). If the shear zone of
the ductile mantle lithosphere controls the deformation, the
thinning of the different layers is vertically superposed,
corresponding to the necking mode of extension (Figure 11c
right).
[41] In experiments at lithospheric scale, the sand layer
which represents the LBL is an isotope level without any
previous fault and the difference of magnitude between the
main shear zones can explain the different evolutions
toward a simple shear or a necking mode. In nature, the
previous history of a tectonic province can control the
formation of structures at lithospheric scale. For example,
an inherited structure in the brittle mantle lithosphere can be
reactivated during continental extension and its geometry
might favor the necking model or the simple shear mode.
[42] To sum up, the extension rate is a key parameter
which controls the geometry of crustal deformation [Michon
and Merle, 2000]. At low extension rate, extension induces
the formation of a single asymmetric graben, whereas
extension leads to the formation of two grabens at high
extension rate. At lithospheric scale, deformation is mainly
controlled by two opposite and main shear zones located
below the asymmetric graben in the lower crust and the
ductile mantle lithosphere. The thinning mode (i.e., simple
shear or necking) is determined by the relative magnitude of
MICHON AND MERLE: MODE OF LITHOSPHERIC EXTENSION
Figure 12. Table showing the interpretation of the
different provinces studied in this paper.
these two main shear zones, which may be due to the
reactivation of inherited structures in the lithosphere. In
both models, the thinning of the lithosphere may induce a
volcanic activity by adiabatic decompression of the mantle.
Volcanism should occur into the asymmetric graben with
the necking mode and across the rift shoulder facing the
master fault in the simple shear mode. These differences in
the continental rift evolution can be considered as evidences
which make possible to decipher the mode which has
occurred in nature.
4.2. Application to Natural Examples
[43] In the first part of this paper, the review of the main
geological features characterizing some continental rift has
shown different evolutions which can exist in nature. We
now apply our model to the natural laboratories in order to
interpret their evolution.
4.2.1. Crustal Geometry
[44] At crustal scale, the Eger graben, the Red Sea rift and
the Rhine graben correspond to single asymmetric grabens
whereas major extensional structures created before oceanization in the North Atlantic passive margins (the Deep
Galicia margin and the East Canadian margin) can be
described as a pair of two companion grabens. We argue
that these two types of basic structures result from a
difference in the extension rate as already proposed by
Michon and Merle [2000] for the Rhine graben and the
Massif Central rift. This suggests that in the Eger graben
and the Red Sea rift where the extension rate during the
early stage of rifting is unknown, the extension rate was
probably lower than in the North Atlantic passive margins
(Figure 12). Using the relation between the strength ratio
and the extension rate [Michon and Merle, 2000], an
extension rate lower than 2 mm/yr can be proposed for
the Eger graben and the first period of extension in the Red
Sea rift. Likewise, an extension rate higher than 2 mm/yr
may have induced the formation of the Deep Galicia margin
and the Jeanne d’Arc basin.
[45] In the different areas (Eger graben, Red Sea rift and
North Atlantic passive margins), the occurrence of one or
two grabens also indicates that the crustal deformation was
caused by a single rupture in the brittle mantle lithosphere.
The Massif Central rift geometry could be explained by (1)
2 - 13
the occurrence of two simultaneous ruptures of the brittle
mantle lithosphere initially spaced by around 50 km and (2)
a high extension rate [Michon and Merle, 2000].
4.2.2. Mode of Extension
[46] In the Red Sea rift and the North Atlantic passive
margins, geological data plead for a simple shear mode with
the development of a main low-angle detachment structure.
In the North Atlantic passive margins, such a mode of
extension allows to explain (1) the predominance of the
reflector S (the crustal shear zone) upon the mantle shear
zone and (2) the lack of mantle lithospheric thinning below
the Jeanne d’Arc basin as it has been shown by Keen and
Dehler [1993]. Thus, the extension was mainly controlled
by a single detachment structure connected with the Deep
Galicia margin, which may have caused the exhumation of
the lithospheric mantle before oceanization. In such a
process, shifting of the extension from the Interior Basin
toward the Deep Galicia margin suggests that the deformation narrows during lithospheric extension.
[47] Following Wernicke [1985], we consider that the
asymmetric geological features resulting from the Late
Oligocene-Early Miocene evolution in the Red Sea rift
can be interpreted as the consequence of the activity of a
single low-angle detachment structure. In contrast, the
present-day data suggest that the mode of lithospheric
extension which is interpreted as the result of a pure shear
mode [Buck et al., 1988] has changed. This indicates that
the mode of extension during continental rifting probably
evoluates toward a mode of extension where the pure shear
deformation prevails during oceanization. At lithospheric
scale, the nearly symmetric shape of opposite passive
margins could therefore result from a change in the mode
of extension after the onset of oceanization.
[48] Assuming that the magmatic activity appears above
the zone of maximum lithospheric thinning, the occurrence
of a syn-rift volcanic phase located within the graben areas
in the Eger graben and the Massif Central rift cannot be
explained by a simple shear mode for which the maximum
of lithospheric thinning is located outside the graben. An
alternative explanation is that the crust is affected by a
single detachment structure, whereas the mantle lithosphere
is thinned in pure shear mode (Figure 3d of Lister et al.
[1991]). However, in this model, the mantle lithosphere is
considered as purely ductile and the resulting strength
profile does not correspond to the strength profile of a
lithosphere with a normal thermal gradient [Davy and
Cobbold, 1991]. In addition, experiments show that once
the lithosphere is characterized by a brittle mantle layer, the
rupture of this brittle part induces the development of shear
zones in both the lower crust and the ductile mantle
lithosphere. This is why we believe that the Eger graben
and the Late Oligocene-Early Miocene evolution of the
Massif Central rift result from lithospheric necking. In these
provinces, necking of the lithosphere has induced a crustal
thinning in the asymmetric graben and the development of
a syn-rift volcanism into the graben. In the Eger graben,
data at crustal scale suggest small stretching values, whereas
the occurrence of the volcanic phase indicates a strong
mantle lithospheric thinning. Such a paradox can be
2 - 14
MICHON AND MERLE: MODE OF LITHOSPHERIC EXTENSION
explained by decoupling the crustal and mantle lithospheric
deformations along the Moho mechanical discontinuity
[Fleitout, 1984].
[49] Although the Rhine Graben likely results from a low
extension rate [Michon and Merle, 2000], the mode of
lithospheric extension cannot be surely constrained with
available geological data as both lithospheric necking and
simple shear mode can generate a single asymmetric graben.
We can only speculate that the reflectors visible below the
Moho and in the continuity of the major fault [Brun et al.,
1992] could represent a part of a continuous detachmentlike structure, which crosscuts the whole lithosphere as in
the simple shear mode.
[50] These conceptual models, based on analogue experiments, indicate that the extension rate and the mode of
extension are the main parameters, which control the
geometry of the continental rifts and passive margins.
Recent numerical experiments suggest that rheological
softening at lithospheric scale can also act on the symmetry
of lithospheric extension, and has a complex interaction
with the rate of extension [Huismans and Beaumont, 2002].
Also, even if our models can explain the main characteristics of several continental rifts, the role of the rheological
distribution at lithospheric scale and sub-surface processes
(i.e., erosion and syn-rift sedimentation) should be considered as potential parameters which can influence the geometry of deformation during continental rifting, as it has been
proposed by Burov and Cloetingh [1997] and Burov and
Poliakov [2001] for post-rift evolution.
[52] As it has already been proposed [Michon and Merle,
2000], the extension rate plays a major role on the crustal
deformation. Narrow rifts (single half-grabens) are formed
with low extension rates whereas wider rift (two grabens)
results from larger extension rate. According to Michon and
Merle [2000], the development of more than two individual
grabens (e.g., the Massif Central rift) results from several
ruptures in the brittle mantle lithosphere.
[53] This study based on analogue experiments also
shows an identical initial rifting stage characterized by
two main shear zones, located below the asymmetric graben
in the lower crust and the ductile mantle lithosphere. With
increasing extension, these two main shear zones may
produce a simple shear mode which shows similarities with
the model proposed by Wernicke [1981, 1985], or a lithospheric necking as described by Brun and Beslier [1996].
These results allow an interpretation of the evolution of the
Rhine graben, the Eger graben, the Massif Central rift, the
Red Sea rift and the North Atlantic passive margins. In
addition, these results could be applied to the North Sea rift
where several mantle lithospheric shear zones have been
identified by seismic studies [Reston, 1993].
[54] Finally, this study suggests an evolution of the mode
of deformation during extension. The continental rifting is
mainly characterized by a simple shear deformation that
induces asymmetric structure in the crust. In contrast, pure
shear deformation probably prevails during oceanization,
explaining the global symmetry of conjugate passive margins superimposed to the asymmetry of the crustal deformation inherited from the continental rift stage.
5. Conclusions
[55] Acknowledgments. This research has been founded by the
ENTEC European project (RTRN-2000-00053). The article is a contribution of the ENTEC and EUCOR-URGENT projects and the IT INSU
‘‘Déformations lithosphériques grandes longueurs d’ondes Cénozoı̈ques de
l’Europe de l’Ouest: Cinématique, Volcanisme, Modélisation’’ French
project. The authors want to thank Muriel Gerbault for useful comments
on an initial version of this manuscript and an anonymous reviewer whose
comments allowed improving the previous version of this manuscript.
[51] This study first shows that two independent sets of
analogue experiments carried out at crustal [Michon and
Merle, 2000] and lithospheric scale [Beslier, 1991; Brun
and Beslier, 1996] can be compared if the models were
initially properly scaled.
References
Allemand, P., J.-P. Brun, P. Davy, and J. Van Den
Driessche, Symetrie et asymetrie des rifts et mecansimes d’amincissement de la lithosphere, Bull. Soc.
Geol. Fr., 5, 445 – 451, 1989.
Baker, J., L. Snee, and M. Menzies, A brief Oligocene
period of flood volcanism in Yemen: Implications
for the duration and rate of continental flood volcanism at the Afro-Arabian triple junction, Earth
Planet. Sci. Lett., 138, 39 – 55, 1996.
Beslier, M. O., Formation des marges passives et remontée du manteau: Modélisation expérimentale et
exemple de la marge de la Galice, Mém. Géosci.
Rennes, 45, 199 pp., 1991.
Beslier, M. O., J. Girardeau, and G. Boillot, Kinematics
of peridotite emplacement during North Atlantic
continental rifting, Galicia, northwestern Spain,
Tectonophysics, 184, 321 – 343, 1990.
Bodergat, A. M., D. Briot, M. Hugueney, J. L. Poidevin,
L. Picot, F. Giraud, J. P. Berger, A. Levy, and A.
Poignant, Incursions marines dans l’environnement
lacustre du rift oligocène de Limagne (Massif central, France): Apport des organismes halophiles et
des isotopes du strontium; datation par les mammifères, Bull. Soc. Géol. Fr., 170, 499 – 511, 1999.
Bohannon, R. G., C. W. Naeser, D. L. Schmidt, and
A. Zimmermann, The timing of uplift, volcanism,
and rifting peripheral to the Red Sea: A case for passive rifting?, J. Geophys. Res., 94, 1683 – 1701, 1989.
Boillot, G., S. Grimaud, A. Mauffret, D. Mougenot,
J. Kornprobst, J. Mergoil-Daniel, and G. Torrent,
Ocean-continent boundary off the Iberian margin:
A serpentinite diapir west of the Galicia bank,
Earth Planet. Sci. Lett., 48, 23 – 34, 1980.
Boillot, G., et al., Proceedings of the Ocean Drilling
Program: Scientific Results, vol. 103, 858 pp.,
Ocean Drill. Program, College Station, Tex., 1988.
Boillot, G., M. O. Beslier, C. M. Krawczyk, D. Rappin,
and T. J. Reston, The formation of passive margins:
Constraints from the crustal structure and segmentation of the deep Galicia margin (Spain), in The
Tectonic, Sedimentation and Palaeoceanography of
the North Atlantic Region, edited by M. S. Stocker,
G. B. Shimmlied, and A. W. Tudhope, Geol. Soc.
Spec. Publ., 90, 71 – 91, 1995.
Bois, C., Initiation and evolution of the Oligo-Miocene
rift basins of southwestern Europe: Contribution of
deep seismic reflection profiling, Tectonophysics,
226, 227 – 252, 1993.
Bosworth, W., Geometry of propagating continental
rifts, Nature, 316, 625 – 627, 1985.
Brown, G. F., D. L. Schmidt, and A. C. Huffman,
Geology of the Arabian Peninsula shield area of
western Saudi Arabia, U.S. Geol. Surv. Open File
Rep., OF-84-0203, 217 pp., 1984.
Brun, J. P., Narrow rifts versus wide rifts: Inferences for
the mechanics of rifting from laboratory experiments, Philos. Trans. R. Soc. London, Ser. A, 357,
695 – 712, 1999.
Brun, J. P., and M. O. Beslier, Mantle exhumation at
passive margin, Earth Planet. Sci. Lett., 142, 161 –
173, 1996.
Brun, J. P., M. A. Gutscher, and DEKORP-ECORS
Teams, Deep crustal structure of the Rhine Graben from DEKORP-ECORS seismic reflexion
data: A summary, Tectonophysics, 208, 139 –
147, 1992.
Buck, W. R., Modes of continental lithospheric extension, J. Geophys. Res., 96, 20,161 – 20,178, 1991.
Buck, W. R., F. Martinez, M. S. Steckler, and J. R.
Cochran, Thermal consequences of lithospheric extension: Pure and simple, Tectonics, 7, 213 – 234,
1988.
MICHON AND MERLE: MODE OF LITHOSPHERIC EXTENSION
Burov, E., and S. Cloetingh, Erosion and rift dynamics:
New thermomechanical aspects of post-rift evolution of extensional basins, Earth Planet. Sci. Lett.,
150, 7 – 26, 1997.
Burov, E., and A. Poliakov, Erosion and rheology controls on synrift and postrift evolution: Verifying old
and new ideas using a fully coupled numerical
model, J. Geophys. Res., 106, 16,461 – 16,481,
2001.
Chazot, G., M. A. Menzies, and J. Baker, Pre-, syn- and
post-rift volcanism on the south-western margin of
the Arabian plate, in Sedimentation and Tectonics of
rift Basins: Red-Sea-Gulf of Aden, edited by B. H.
Purser and D. W. J. Bosence, pp. 50 – 55, Chapman
and Hall, New York, 1998.
Chemenda, A., J. Deverchere, and E. Calais, Threedimensional laboratory modelling of rifting: Application to the Baikal rift, Russia, Tectonophysics,
356, 253 – 273, 2002.
Chlupac, I., O. Kodym, M. Suk, V. Holub, M. Elias,
I. Cicha, M. Malkovsky, J. Tyracek, E. Mencyk, and
T. Buday, Stratigraphic development of the units, in
Geological History of the Territory of the Czech
Socialist Republic, edited by M. Suk et al., pp.
87 – 200, Geol. Surv., Prague, 1984.
Cochran, J. R., and F. Martinez, Evidence from the
northern Red Sea on the transition from continental
to oceanic rifting, Tectonophysics, 153, 25 – 53,
1988.
Courtillot, V. E., Opening of the Gulf of Aden and Afar
by progressive tearing, Phys. Earth Planet. Inter.,
21, 343 – 350, 1980.
Davy, P., and P. R. Cobbold, Experiments on shortening
of a 4-layer model of the continental lithosphere,
Tectonophysics, 188, 1 – 25, 1991.
Dixon, T. H., E. R. Ivins, and B. J. Franhlin, Topographic and volcanic asymmetry around the Red
Sea: Constraints on rift models, Tectonics, 8,
1193 – 1216, 1989.
Doebl, F., and W. Olbrecht, An isobath map of the
Tertiary base in the Rhine graben, in Approaches
to Taphrogenesis: Proceedings of an International
rift Symposium Held in Karlsruhe, edited by J. H.
Illies and K. Fuchs, pp. 71 – 72, Schweizerbart,
Stuttgart, Germany, 1974.
Driscoll, N. W., J. R. Hogg, N. Cristie-Blick, and
G. Karner, Extensional tectonics in the Jeanne
d’Arc Basin, offshore Newfoundland: Implications
for the timing of the break-up between Grand
Banks and Iberia, in The Tectonic, Sedimentation
and Palaeoceanography of the North Atlantic Region, edited by M. S. Stocker, G. B. Shimmlied,
and A. W. Tudhope, Geol. Soc. Spec. Publ., 90,
1 – 28, 1995.
Faugère, E., J. P. Brun, and J. van dan Driessche,
Bassins asymétriques en extension pure et en
décrochement: Modèles expérimentaux, Bull. Cent.
Rech. Explor. Prod. Elf Aquitaine, 10, 13 – 21,
1986.
Féraud, G., J. Girardeau, M. O. Beslier, and G. Boillot,
Datation 39Ar/40Ar de la mise en place des peridotites bordant la marge de la Galice (Espagne), C. R.
Acad. Sci. Paris, 307, 49 – 55, 1988.
Fernandez, M., and G. Ranalli, The role of rheology in
extensional basin formation modeling, Tectonophysics, 282, 129 – 145, 1997.
Fleitout, L., Modélisation des contraintes tectoniques et
des instabilités thermo-mécaniques dans la lithosphère, Thèse d’état, 433 pp., Univ. Orsay, France,
1984.
Girardeau, J., C. A. Evans, and M. O. Beslier, Structural analysis of plagioclase-bearing peridotites emplaced at the end of continental rifting: Hole
637A, ODP Leg 103 on Galicia margin, Proc.
Ocean Drill. Program Sci. Results, 103, 209 –
223, 1988.
Giraud, J., Etudes géologiques sur la Limagne (Auvergne), Thèse d’état, 410 pp., Ed. Ch. Béranger,
Paris, 1902.
Granet, M., S. Judenherc, and A. Souriau, Des images
du système lithosphère-asthénosphère sous la
France et leurs implications géodynamiques: L’apport de la tomographie télésismique et de l’anisotropie sismique, Bull. Soc. Géol. Fr., 171, 149 –
167, 2000.
Hubbert, K. M., Theory of scale models as applied to
the study of geologic structures, Geol. Soc. Am.
Bull., 48, 1459 – 1520, 1937.
Huismans, R. S., and C. Beaumont, Asymmetric lithospheric extension: The role of frictional plastic
strain softening inferred from numerical experiments, Geology, 30, 211 – 214, 2002.
Jaeger, J. C., and N. G. W. Cook, Fundamental of Rock
Mechanics, 585 pp., Chapman and Hall, New York,
1971.
Keen, C. E., and S. A. Dehler, Stretching and subsidence: rifting of conjugate margins in the
North Atlantic Region, Tectonics, 12, 1209 –
1229, 1993.
Kooi, H., M. Hettema, and S. Cloetingh, Lithospheric
dynamics and the rapid Pliocene-Quaternary subsidence phase in the Southern North Sea Basin, Tectonophysics, 192, 245 – 259, 1991.
Kusznir, N. J., and P. A. Ziegler, The mechanics of
continental extension and sedimentary basin formation: A simple-shear/pure-shear flexural cantilever model, Tectonophysics, 215, 117 – 131,
1992.
Le Pichon, X., and J. M. Gaulier, The rotation of Arabia
and the Levant Fault System, Tectonophysics, 153,
271 – 294, 1988.
Lister, G. S., M. A. Etheridge, and P. A. Symonds,
Detachment models for the formation of passive
continental margins, Tectonics, 10, 1038 – 1064,
1991.
Manatschal, G., and D. Bernoulli, Architecture and tectonics evolution of nonvolcanic margins: Presentday Galicia and ancient Adria, Tectonics, 18,
1099 – 1119, 1999.
Mart, Y., and J. K. Hall, Structural trends in the northern Red Sea, J. Geophys. Res., 89, 11,352 – 11,354,
1984.
Mauffret, A., and L. Montadert, Rift tectonics on the
passive continental margins off Galicia (Spain),
Mar. Pet. Geol., 4, 49 – 70, 1987.
McKenzie, D., Some remarks on the development of
sedimentary basins, Earth Planet. Sci. Lett., 40,
25 – 42, 1978.
Menzies, M., J. Baker, G. Chazot, and M. Al’Kadasi,
Evolution of the Red sea volcanic margin, Western
Yemen: Large igneous provinces, in Continental,
Oceanic, and Planetary Flood Volcanism, Geophys.
Monogr. Ser., vol. 100, edited by J. Mahoney and
M. Coffin, pp. 29 – 43, AGU, Washington, D. C.,
1997.
Merle, O., and A. Borgia, Scaled experiments of volcanic spreading, J. Geophys. Res., 101, 13,805 –
13,817, 1996.
Merle, O., L. Michon, and G. Camus, L’extension oligocène sur la transversale septentrionale du rift du
Massif Central, Bull. Soc. Géol. Fr., 169, 615 – 626,
1998.
Merle, O., N. Vidal, and B. Van Wyk de Vries, Experiments on vertical basement fault reactivation below
volcanoes, J. Geophys. Res., 106, 2153 – 2162,
2001.
Michon, L., Dynamique de l’extension continentaleApplication au rift Ouest-Européen par l’étude de
2 - 15
la province du Massif Central, Mém. Géosci.
Rennes, 99, 266 pp., 2001.
Michon, L., and O. Merle, Crustal structures of the
Rhinegraben and the Massif Central grabens: An
experimental approach, Tectonics, 19, 896 – 904,
2000.
Ranalli, G., and D. C. Murphy, Rheological stratification of the lithosphere, Tectonophysics, 132, 281 –
296, 1987.
Rat, P., Le système Bougogne-Morvan-Bresse (articulation entre le bassin parisien et le domaine périalpin), in Géologie de la France: Les Chaı̂nes
Plissées du Cycle Alpin et Leur Avant-Pays, edited
by J. Debelmas, pp. 480 – 500, Doin, Paris, 1974.
Reston, T. J., Evidence for extensional shear zones in
the mantle, offshore Britain, and their implications
for the extension of the continental lithosphere, Tectonics, 12, 492 – 506, 1993.
Reston, J. T., C. M. Krawczyk, and D. Klaeschen, The
S reflector west of Galicia (Spain): Evidence from
prestack depth migration for detachment faulting
during continental breakup, J. Geophys. Res., 101,
8075 – 8091, 1996.
Rittmann, U., and H. J. Lippolt, Evidence for distortion
of Tertiary K/Ar ages by excess argon-example given by three alkali olivine basalts from Northern
Hesse, Germany, Eur. J. Mineral., 10, 95 – 110,
1998.
Rosendahl, B. R., E. Kilembe, and K. Kaczmarick,
Comparison of Tanganyika, Malawi, Rukwa and
Turkana rift zones from analyses of seismic reflection data, Tectonophysics, 213, 235 – 256, 1992.
Schumacher, M., Upper Rhine Graben: Role of preexisting structures during rift evolution, Tectonics,
21(1), 1006, doi:10.1029/2001TC900022, 2002.
Sissingh, W., Comparative Tertiary stratigraphy of the
Rhine Graben, Bresse Graben and Molasse Basin:
Correlation of Alpine foreland events, Tectonophysics, 300, 249 – 284, 1998.
Tankard, A. J., and H. J. Welsink, Extensional tectonics
and stratigraphy of Hiberia Oil Field, Grand Banks,
Newfoundland, AAPG Bull., 71, 1210 – 1232, 1987.
Ulrych, J., V. Cajz, E. Pivec, J. K. Novak, C. Nekovarik,
and K. Balogh, Cenozoic intraplate alkaline volcanism of western Bohemia, Stud. Geoph. Geod., 44,
346 – 351, 2000.
Villemin, T., F. Alvarez, and J. Angelier, The Rhinegraben: Extension subsidence and shoulder uplift,
Tectonophysics, 128, 47 – 59, 1986.
Vissers, R. L. M., M. R. Drury, E. H. Hogerduijn Strating, C. J. Spiers, and D. van der Wal, Mantle shear
zones and their effect on lithosphere strength during
continental breakup, Tectonophysics, 249, 155 –
171, 1995.
Wernicke, B., Low-angle normal faults in the Basin and
Range Province: Nappe tectonics in an extending
orogen, Nature, 291, 645 – 648, 1981.
Wernicke, B., Uniform-sense normal simple shear of
the continental lithosphere, Can. J. Earth Sci., 22,
108 – 125, 1985.
Zijerveld, L., R. Stephenson, S. Cloetingh, E. Duin, and
M. W. van den Berg, Subsidence analysis and modelling of the Roer Valley Graben (SE Netherland),
Tectonophysics, 208, 159 – 171, 1992.
O.
Merle,
Laboratoire Magmas et Volcans, OPGC,
Université Blaise Pascal, CNRS, F-63038 ClermontFerrand cedex, France.
L. Michon, Netherlands Organization for Applied
Scientific Research – Netherlands Institute of Applied
Geoscience (TNO-NITG), P.O. Box 80015, NL-3508
TA Utrecht, Netherlands. ([email protected])
MICHON AND MERLE: MODE OF LITHOSPHERIC EXTENSION
Figure 7. Internal strain of the silicone layers in high extension rate experiments deduced from the final
geometry of the vertical markers in the silicone layers. Colors from light to dark blue indicate variations
in strain intensity for the top-to-the-left shear zone whereas the yellow to red colors indicate strain
intensity for the top-to-the-right shear zone (see text for explanation).
2-9